Since endocrinologists proved they could name studies equally as well as cardiologists with the RABBIT 2 and NICE SUGAR trials, guidelines have recommended the use of insulin for inpatients with type 2 diabetes to maintain a blood sugar between 140 and 180 mg/dL for most patients due to the ease of titration and predictable pharmacokinetics. Oral agents have been avoided as they are difficult to titrate and have side effects that could be exacerbated in inpatients who have variable oral intake, are experiencing acute illness, have renal and hepatic dysfunction, are undergoing imaging procedures, or are in the peri-operative phase and face an increased risk.
With newer anti-diabetic agents being used and often started as inpatients for other indications, in this Flipside, we argue over the use of non-insulin diabetic agents in the hospital.
Limited Unpredictability with Insulin Therapy Alone
It’s my first day back on a busy hospitalist service. I’ve settled in with my morning coffee and started to review my patients when a rapid response is called overhead to room 764. Looking down at my list and recognizing the patient that was signed out to me with type 2 diabetes mellitus, chronic systolic congestive heart failure (CHF), mild acute kidney injury (AKI), and advanced peripheral artery disease (PAD), who is on the schedule for peripheral intervention later that day, I quickly get up and rush to the elevator.
Upon arriving in the room, I see a toxic-appearing 55-year-old male, diaphoretic, tachypneic, lethargic, and largely obtunded. Vital checks reveal a blood pressure of 108/67, a heart rate of 123, a respiratory rate of 28, O₂ saturation of 94% on room air, and a temperature of 98.3 °F. His morning labs have yet to return, and his nurse is visibly shaken because the patient was “fine” when he went to sleep the prior evening. A glucose is checked: 188 mg/dL. Given this unexpected and undifferentiated decompensation, I order a transfer to a higher level of care for a more thorough workup.
As I dig into the chart while the patient is being moved, I notice he has had his peripheral intervention rescheduled twice this week due to OR availability. His medications include sliding scale insulin and continued dapagliflozin on admission reconciliation. Shortly after, his comprehensive metabolic panel (CMP) returns with a high-anion-gap metabolic acidosis and a CO2 of 11. Recognizing that prolonged nil per os status combined with his SGLT2 inhibitor had pushed him into euglycemic diabetic ketoacidosis (DKA), I start 2 liters of IV fluids and insulin and dextrose infusions. By the afternoon, his acidosis and lethargy have resolved, but his peripheral stent has been pushed back another week by a nervous surgeon.
Good hospital medicine practice has always been about reducing unpredictability in an inherently unpredictable environment. Glycemic management is one of those areas where we can exert control, and basal-bolus insulin therapy has long been the tool that allows us to do so safely. Oral agents, by contrast, add unnecessary uncertainty. Inpatients often have variable or unreliable oral intake, are kept nil per os for procedures, or become acutely ill in ways that change their nutritional intake overnight. In that context, medications designed for stable outpatients can precipitate dangerous hypoglycemia, and in this patient, euglycemic DKA. The juice of continuing these medications is generally not worth the squeeze.
Renal and hepatic dysfunction are frequent companions of hospitalization, making medications like metformin and sulfonylureas particularly hazardous. What works at home in a stable setting can suddenly become toxic in the hospital, where AKI or hypoperfusion may push a patient into lactic acidosis or prolonged hypoglycemia. Add to this the common exposures to contrast, high-dose steroids, or perioperative metabolic shifts, and the risks of continuing oral therapy quickly add up.
Decrease the Sticks and Use Oral Agents in Certain Circumstances
My sixth admission comes into the hospital, and before truly digging through my chart, I’ve already placed diabetes as a problem on my problem list and have my inpatient insulin order set pre-clicked to make sure at least Accu-Chek, sliding scale insulin, and hypoglycemia protocols are ordered. Yesterday, I overheard nurses complaining about drawing up one unit of insulin for a blood sugar of 185 mg/dL, and the patient was annoyed at another stick for such a small amount of medication. He argued with the nurses that he’s not on insulin at home, so why don’t we just continue his home medications?
I think about the recent heart failure patient whom I inherited, who had their dapagliflozin continued when they were re-admitted after their new diagnosis, where guideline-directed medical therapy (GDMT) was started. I monitored him for DKA; he had no issues, and he needed less insulin. A colleague mentioned that they routinely use sitagliptin where he practiced to reduce the amount of sliding scale. Admission six was stable, his kidneys were working well, and he didn’t get any contrast. What is the harm in continuing his metformin? If he develops lactic acidosis, I can just stop it, and currently, he’s at such low risk. I also think about how, with sliding scale, I’m constantly chasing the sugar. Administering medication after the patient has eaten can lead to complications, especially when considering the post-prandial check and the late timing of the sliding scale. I often notice that guideline-based basal-bolus recommendations are not followed, resulting in patients relying on a sliding scale throughout their hospitalization, with blood sugar levels frequently remaining in the 300s.
Coming back to admission six, a patient with a mild cellulitis and normal kidneys, and who is not getting nephrotoxic meds or imaging, I continue his home oral regimen of metformin, linagliptin, and dapagliflozin. His sugars stay okay, as does his renal function, and he doesn’t require any doses of sliding scale. He is happy; the nurses are happy. For this low-risk patient, where unpredictability is more limited and he’s being actively monitored for side effects, oral meds were a safe and decent choice. It reduced the amount of wasted medications, the number of needle sticks, and avoided some of the risks of fluctuations that sliding scale can cause. Now he wants to just use his continuous glucose monitoring device instead to completely avoid the needles…a debate we can have another time.
Discussion
While insulin should remain the mainstay of therapy, as new evidence and clinical practice changes, our inpatient 2,3 practice should adapt. Guidelines for inpatient management of diabetes have recently been updated to include the use of dipeptidyl peptidase 4 (DPP4) inhibitors.1 Strong, randomized, controlled trials have shown the efficacy of the use of sitagliptin and linagliptin for inpatients for better glycemic control and also for less overall insulin need, which may mean a few fewer sticks for our inpatients.2,3 Clinically, they can be used for mild-to-moderate hyperglycemia in stable, noncritically ill patients with type 2 diabetes mellitus. They may be most beneficial when we have those folks in between 180 and 200 mg/ dL, and our nurses are using one unit of sliding scale to cover their mild hyperglycemia. Consistent oral intake is still important, and if sitagliptin is used, it has to be renally adjusted.
SGLT2 inhibitors are being used clinically. With so much evidence toward starting these medications for our patients with heart failure as inpatients, we are now seeing these folks when they come back to the hospital. There isn’t great evidence on what to do with these agents in terms of randomized control trials, but in practice, they are being used and continued. A nationwide cohort study using Veterans Affairs healthcare system data showed some benefits of continuing these agents.4 However, the risk of euglycemic DKA is real, and they should be stopped peri-operatively or during periods of prolonged poor oral intake. Additionally, they should be stopped in patients with acute urinary infections or pyelonephritis. Volume status may be another consideration for holding.
Sulfonylureas and thiazolidinediones still don’t have a role in inpatient medicine. However, metformin, our tried-and-true oral medication, is one that is generally not continued as an inpatient. Our fellow hospitalists have argued that it is one of the Things We Do For No Reason™ and that there are scenarios where metformin can be used, especially if a patient has stabilized and is nearing the end of discharge.5 The risk of metformin-associated lactic acidosis may be exaggerated from the proguanil days, but still should be avoided in patients with sepsis, renal failure, recent contrast imaging, or other reasons to be at high risk for lactic acidosis.
GLP-1s are now commonplace as well; typically, they are not on formulary and are somewhat impractical for inpatient use given the pen delivery system. However, GLP-1s are something for hospitalists to think about on discharge for patients with obesity or those who were admitted for a stroke, with their proven cardiovascular benefits.6
There may be a role for oral diabetic agents in select patients, and it should be individualized based on the clinical setting and patient. However, blindly continuing SGLT2 inhibitors is not advised due to the risks associated with them. The role of DPP4 inhibitors has good data and should be considered in non-critically ill patients, while metformin can provoke a continued argument over whether we hold it for no reason. When patients have added unpredictability in their renal and hepatic function, perfusion, or oral intake status, sticking with insulin is the safest approach.
Dr. Molitch-Hou
Dr. Donohue
Dr. Molitch-Hou is an assistant professor, the director of the hospital medicine sub-internship, core faculty for the internal medicine residency program, and co-director of the Care Transition Clinic at the University of Chicago Medical Center in Chicago. Dr. Donohue is a practicing hospitalist and regional medical director for Team Health as well as the chief of medicine at Georgetown Community Hospital in Georgetown, Ky. He serves as the medical director of telemedicine for Team Health’s Strategic Account Group and is a member of SHM’s Community Hospitalist Advisory Board.
References
1. American Diabetes Association Professional Practice Committee. 16. Diabetes care in the hospital: standards of care in diabetes-2025. Diabetes Care. 2025;48(1 Suppl 1):S321-S334. doi: 10.2337/dc25-S016.
2. Vellanki P, et al. Glycaemic efficacy and safety of linagliptin compared to a basal-bolus insulin regimen in patients with type 2 diabetes undergoing non-cardiac surgery: A multicentre randomized clinical trial. Diabetes Obes Metab. 2019;21(4):837- 843. doi: 10.1111/dom.13587.
3. Pasquel FJ, et al. Efficacy of sitagliptin for the hospital management of general medicine and surgery patients with type 2 diabetes (Sita-Hospital): a multicentre, prospective, open-label, non-inferiority randomised trial. Lancet Diabetes Endocrinol. 2017;5(2):125-133. doi: 10.1016/S2213- 8587(16)30402-8.
4. Singh LG, et al. Association of continued use of SGLT2 Inhibitors from the ambulatory to inpatient setting with hospital outcomes in patients with diabetes: a nationwide cohort study. Diabetes Care. 2024;47(6):933-940. doi: 10.2337/dc23-1129.
5. Cohen DA, et al. Things we do for no reason™: routinely holding metformin in the hospital. J Hosp Med. 2022;17(3):207- 210. doi: 10.12788/jhm.3644.
6. Adamou A, et al. Glucagon-like peptide-1 receptor agonists and stroke: a systematic review and meta-analysis of cardiovascular outcome trials. Int J Stroke. 2024;19(8):876-887. doi: 10.1177/17474930241253988.
